1. Study Title: Social adjustment in adolescent survivors of pediatric CNS tumors 1 2. Working Group and Investigators: Working Group: Psychology Investigators: Fiona Schulte Fiona.schulte@ahs.ca Kevin R. Krull (Mentor) Kevin.Krull@stjude.org Tara Brinkman tara.brinkman@stjude.org Wendy Leisenring wleisenr@fhcrc.org Greg Armstrong greg.armstrong@stjude.org Les Robison les.robison@stjude.org Kiri Ness kiri.ness@stjude.org Rebecca Howell rhowell@mdanderson.org Sabine Mueller muellers@peds.ucsf.edu Elizabeth Wells ewells@childrensnational.org Taryn Fay-McClymont taryn.faymcclymont@ahs.ca Doug Strother doug.strother@ahs.ca Lucie Lafay-Cousin lucie.lafaycousin@ahs.ca 3. Background and Rationale Improved therapies for pediatric CNS tumor have increased survival rates, however, many survivors experience significant long-term functional limitations. CNS tumor survivors have been found to experience deficits in social adjustment (i.e., the ability to achieve personal goals in social interactions while maintaining positive relationships with others over time and across situations 1 ) that worsen with time, 2 and negatively affect survivors long-term quality of life. 3 Currently there are more than 115,000 survivors of pediatric central nervous system (CNS) tumor in North America. 4,5 As adults, these survivors are more likely than their siblings to require special education services, are less likely to attend college, less likely to live independently, and are at increased risk of unemployment 6. In addition, survivors of pediatric CNS tumors are less likely to be married compared to their siblings 6. However, over the past several decades research in the social adjustment of pediatric CNS tumor survivors has been limited, relying on small, single center cohort studies (average sample size = 66 2 ) which has remained the primary criticism of research conducted in the field, to date 2,7. Although there are a few notable studies that have been completed using large sample sizes, one in particular using the Childhood Cancer Survivor Study (CCSS) cohort 8, a detailed examination of risk and resilience factors specific to CNS tumor survivors and examination of specific social problems has not been completed. There is a paucity of research directly investigating the type and severity of CNS insult and associations with social adjustment in pediatric CNS tumor survivors. 9 Where research has attempted to explore the impact of disease and treatment factors on social adjustment, little conclusive evidence has been obtained. 10 12 For example, cranial radiation therapy (CRT) and young age at diagnosis have been identified as risk factors for cognitive deficits among pediatric CNS tumor survivors, 13 however, there is less evidence to support the role of CRT in poor social adjustment. 8,14,15 The evidence related to age at diagnosis is mixed. 12,14 Further research is needed with a large diverse sample to better identify predictors of long-term social adjustment difficulties in this population. 11 Demographic factors, such as sex and age have yielded contradictory results with respect to their role in social adjustment in pediatric brain tumor survivors. 16,17 In children with mild traumatic brain injury, male gender and younger age has been found to contribute to poorer social adjustment outcomes. 18 Socio-economic status (SES) plays an established role in social development for typically developing children 19 and children with other acquired brain injuries (i.e., traumatic brain injury[tbi]) 20 whereby lower SES is linked to poorer social outcomes. There is no research that has directly examined the relationship of SES on social adjustment outcomes in pediatric CNS tumor survivors. Poor physical mobility is often seen in survivors of pediatric CNS tumors, and may further compromise social adjustment outcomes. 6,21 Recent research has revealed that as adults, reduced physical function in survivors of pediatric CNS tumor has been linked to poor environmental access which was associated with reduced quality of life and social
function. 22 The proposed study will be the first to examine potential moderation effects of physical limitations on social adjustment in an adolescent sample of pediatric CNS tumor survivors. 2 We propose to examine social function in a large sample of CNS tumor survivors from the 1970-99 CCSS cohort. We will focus on those survivors whose parents (proxy) completed the Baseline surveys when the survivor was < 18 years of age. Social function in this group will be compared to a non-cns solid tumor comparison group (i.e., neuroblastoma, Wilms tumor) as well as siblings whose parents (proxy) participated in the < 18 Baseline survey. 4. Specific Aims Aim 1: To examine patterns of social adjustment (e.g., number of close friends, frequency of interactions, quality of interactions, social withdrawal, conflict) in adolescent survivors of pediatric CNS tumors in the combined cohort as compared to a non-cns tumor comparison group (i.e., patients diagnosed with neuroblastoma, Wilms tumor) and to the sibling cohort. Hypothesis 1: We hypothesize that pediatric brain tumor survivors will have fewer friends, fewer interactions, poorer quality of social interactions and greater social withdrawal and conflict compared to non-cns tumor survivors and siblings. Aim 2: To identify demographic (e.g., sex, age), socioeconomic (e.g., household income), disease (e.g., tumor diagnosis), and treatment (e.g., radiation dose and site, age at diagnosis) factors that are related to social function in adolescent survivors of pediatric CNS tumor. Hypothesis 2: Male gender, lower household income, diagnosis of Medulloblastoma and/or PNET, higher dose of cranial radiation, and younger age at diagnosis will be associated with poorer social adjustment outcomes. Aim 3: To examine the association between and moderation effect of physical, cognitive and sensory limitations (e.g., weakness/paralysis, poor endurance, special education services and vision and/or hearing loss) and social adjustment in adolescent survivors of pediatric CNS tumor. Hypothesis 3: Physical, cognitive and sensory limitations will moderate the effects of personal demographic, socioeconomic, disease and treatment factors on social adjustment outcomes. 5. Analysis Framework Population Participants will be adolescent survivors of pediatric CNS tumors in the original and expansion cohort for whom the baseline survey was completed by their parent (proxy) when they were between the ages of 12 to 17 years, and siblings for whom the baseline survey was completed by their parent (proxy) when they were 12 to 17 years of age. There are 663 survivors of CNS tumors in the combined cohort, 1374 non-cns solid tumor survivors in the combined cohort, and 649 siblings in the original cohort who meet these inclusion criteria. Note: we will also include siblings < 18 years at Baseline from the expansion cohort, if the data is available at the time of analyses. Variables Variables for Aim 1 Social Adjustment: measured with the social withdrawal scale from the Behavior Problems Index (BPI) 23 and individual questions in the Baseline < 18 survey (Original; Expansion): Number of Close Friends (J.16; K.1), Frequency of Interactions (J.17; K.2), Quality of Interactions (J.18 a,b,c,d; K.3 a,b,c,d), Social Withdrawal/Peer Conflict (J.19 m,q,w; K.4 m.q.w), Antisocial behavior (J.19 d,i,l,j,n; K.4 d,i.l,j,n).
3 The BPI 23 is a 28-item scale that is completed by proxy reporters for adolescents 12-17 years of age with responses recorded on a three-point Likert scale ranging from not true to often true. The items that comprise the BPI are a subset of items from the Child Behavior Checklist (CBCL). 24 Internal reliability estimates for the BPI have been reported as α = 0.86 for peer conflict/social withdrawal. 8 Variables for Aim 2 Outcome Variables Social Adjustment (Original; Expansion): Number of Close Friends (J.16; K.1), Frequency of Interactions (J.17; K.2), Quality of Interactions (J.18 a,b,c,d; K.3 a,b,c,d), Social Withdrawal/Peer Conflict (J.19 m,q,w; K.4 m.q.w), Antisocial behavior (J.19 d,i,l,j,n; K.4 d,i.l,j,n). Predictors Personal Demographic o o Sex (A.2; male or female). Current age (Original; Expansion). Socioeconomic Status o Household Income (Q.8; T.1; Original, Expansion). Income will be categorized as: 1) <$20,000; 2) $20,000-$39,999; 3) $40,000-$59,999; and 4) $60,000. o Family size (A.7; Baseline; Expansion). Disease o Tumor Diagnosis will be operationalized using four diagnosis categories: 1) Astrocytoma; 2) Medulloblastoma; PNET; 3) Ependymoma and 4) Other CNS tumor. Treatment o Radiation dosimetry will be operationalized consistent with research that has previously examined the effect of radiation on psychological/neuropsychological outcomes 25. Specifically, we will use four segments of brain including: 1) posterior fossa ; 2) temporal lobe; 3) frontal cortex; and 4) parietal or occipital lobe. Using data abstracted from the radiation therapy records, average and maximum radiation dose to each segment of the brain was previously calculated. For each case, each of the four brain segments will be assigned to one of the following dose categories based on average dose to the segment 1) none; 2) 0.1-29 Gy; 2) 30-49 Gy; and 3) >50 Gy. o Age at diagnosis. Variables for Aim 3 Social Adjustment (Baseline; Expansion): Number of Close Friends (J.16; K.1), Frequency of Interactions (J.17; K.2), Quality of Interactions (J.18 a,b,c,d; K.3 a,b,c,d), Social Withdrawal/Peer Problems (J.19 m,q,w; K.4 m.q.w), Conflict/antisocial (J.19 d,i,l,j,n; K.4 d,i.l,j,n). Physical limitations: as measured by the SF-36, Role-Limitations - Physical (N. 10 a,b,c,d,e,f; O. 6 a,b,c,d,e,f; Original; Expansion). Cognitive Limitations: as measured by School History (Baseline; Expansion): O3; R3. Sensory impairments as measured by Hearing/Vision/Speech (Baseline; Expansion): (C.1, C.2, C.3, C.8; C.1, C.2, C.3, C.8, C.9) Statistical Analyses
Descriptive Characteristics of the Sample 4 Demographic characteristics of the sample (e.g., age, gender of child, gender of parent, socioeconomic variables) will be calculated and comparisons between survivors and siblings will be made using t-tests or chi-square where appropriate. Potential impact of the proxy-reporter (i.e., mother, father, guardian) will also be considered and if differences exist, controlled for in subsequent analyses. Statistical Analysis for Aim 1 Social adjustment will be operationalized using the five items specified above (i.e., number of close friends (categorical, based on 4 responses), frequency of Interactions (categorical based on 3 responses), quality of interactions (categorical based on 4 responses), social withdrawal/peer problems (continuous based on 3 items), conflict/antisocial behavior) (continuous based on 4 items). Distributions will be examined to identify categories of sufficient size for each variable that represent poor outcomes (e.g. <1 or <2 close friends). Latent profile analysis (LPA) will be used to identify clusters of siblings based on item level responses of social adjustment variables. Latent clusters will first be identified in the sibling cohort (discovery) and validated in the survivor cohort to verify the same general cluster pattern. For the discovery set, the number of potential classes will not be prespecified. A minimum of 5% of the sample will be required in each cluster to be included. The centers of the cluster model in the sibling cohort will then be used to derive the latent clusters in the survivor cohort using nearest centroid method. Frequencies and percentages of cluster membership will be compared between survivors and siblings using Chisquare tests adjusting for those variables for which they differ (e.g., gender, household income, family size). Statistical Analysis for Aim 2 Utilizing the same latent clusters(s) for social adjustment derived in Aim 1, multivariable logistic regression analyses will be conducted for each cluster as a dichotomized variable (determined based on class membership yes/no). That is, membership in a given cluster will be predicted using the following variables: tumor diagnosis; radiation dosimetry; age at diagnosis; sex; age; household income. Separate models will be run for disease and treatment variables to avoid confounding. Statistical Analysis for Aim 3 Multivariable logistic regression will also be used to examine associations between physical limitations/cognitive limitations/sensory impairments and clusters of social adjustment outcomes. To examine the potential moderation effect of physical limitations/cognitive limitations/sensory impairments on social adjustment we will: 1) identify the significant relations between the independent variables (i.e., tumor diagnosis; radiation dosimetry; age at diagnosis; gender; age; household income; and family size) and social adjustment as the dependent variable(s) assessed in Aim 2; 2) complete three separate multivariable regression analyses that examine whether the independent variables (i.e., tumor diagnosis; radiation dosimetry; age at diagnosis; gender; age; household income; and family size) significantly predict: 1) physical limitations; 2) cognitive limitations; 3)sensory impairments; and finally 3) complete three separate hierarchical regressions to examine whether: 1) physical limitations; 2) cognitive limitations; and 3) sensory impairments are significant predictors of social adjustment, while controlling for the independent variables. We will evaluate moderation using the methods described by Baron and Kenny 26 6. Tables Table 1. Participant Characteristics
Age Sex Family Income Ethnicity Race Age at Diagnosis Time Since Diagnosis Diagnosis Male Female <$20,000 $20,000 - $39,999 $40,000 - $59,999 $60,000 Hispanic Non-Hispanic White Black Other Astrocytoma Medulloblastoma, PNET Other CNS Tumor Cranial Radiation None Whole Brain Partial Brain Cranial Radiation Dose None 0.1-29 Gy 30-59 Gy >50Gy CNS Survivor (n =) Non-CNS Solid Tumor Survivor (n = ) Sibling (n =) N (%) M (SD) N (%) M (SD) p 5
Table 2. Model fit indices for two to six social cluster solutions from the datasets of siblings and survivors 6 Siblings Survivors Class 2 3 4 5 6 2 3 4 5 6 AIC BIC abic Entropy LMR test Class % Table 3. Comparison of Social Adjustment Among Survivors and Siblings
Total (n / %) Astrocytoma (n / %) CNS Survivors Medulloblastoma, PNET (n / %) Other (n / %) Non-CNS Solid Tumor Survivors (n/%) Siblings (n / %) 7 P (CNS vs. Siblings) p ( Number of Close Friends Frequency of Interactions 0 1 2 or 3 4 or more Less than 1 1 or 2 3 or more Quality of Interactions Get along with his/her brothers and sisters Get along with other children Behave with his/her parents Play and work by himself/herself Social Withdrawal/Peer Problems Has trouble getting along with other children Is not liked by other children Is withdrawn, does not get involved with others Conflict/Antisocial Cheats or tells lies Bullies, or is cruel or mean to others Does not seem to feel sorry after he/she misbehaves Is disobedient at home Has trouble getting along with teachers
Table 4a. Multivariable Logistic Regression Analysis of Factors Predicting Social Adjustment 8 Variables Social Class 1 Social Class 2 OR CI 95% OR CI95% Personal Demographic Age (per year) Sex Female 1.0 1.0 Male Sociodemographic Household Income* <$60,000 $60,000 1.0 1.0 Diagnosis Astrocytoma Medulloblastoma, PNET Other CNS Tumor 1.0 1.0 Treatment Age at Diagnosis (per year) *Adjusted to Expansion cohort values. Table 4ba. Multivariable Logistic Regression Analysis of Factors Predicting Social Adjustment Variables Social Class 1 Social Class 2 OR CI 95% OR CI95% Personal Demographic Age (per year) Sex Female 1.0 1.0 Male Sociodemographic Household Income* <$60,000 $60,000 1.0 1.0 Treatment Age at Diagnosis (per year) Cranial Radiation Dose None 1.0 1.0 0.1-29 Gy 30-49 Gy >50Gy *Adjusted to Expansion cohort values.
Figure 1. Examination of the moderation effect of physical limitations on social adjustment 9 Figure 2. Examination of the moderation effect of cognitive limitations on social adjustment. Figure 3. Examination of the moderation effect of sensory impairments on social adjustment.. References
10 1. Yeates KO, Bigler ED, Dennis M, et al. Social outcomes in childhood brain disorder: A heuristic integration of social neuroscience and developmental psychology. Psychol Bull. 2007;133:535 556. 2. Schulte F, Barrera M. Social competence in childhood brain tumor survivors: a comprehensive review. Support Care Cancer. 2010;18:1499 1513. 3. Zebrack B, Gurney J, Oeffinger K, et al. Psychological outcomes in long-term survivors of childhood brain cancer: A report from the childhood cancer survivor study. J Clin Oncol. 2004;22:999 1006. 4. Brain Tumour Foundation of Canada. No Title. Available at: http://www.braintumour.ca/2494/brain-tumour-facts. Accessed September 20, 2015. 5. Ward E, DeSantis C, Robbins A, Kohler B, Jemal A. Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin. 64(2):83 103. 6. Gurney JG, Krull KR, Kadan-Lottick N, et al. Social outcomes in the Childhood Cancer Survivor Study cohort. J Clin Oncol. 2009;27(14):2390 5. 7. Schulte F. Social competence in pediatric brain tumor survivors: breadth versus depth. Curr Opin Oncol. 2015;27(4):306 10. 8. Hocking M, McCurdy M. Social competence in pediatric brain tumor survivors: Application of a model from social neuroscience and developmental psychology. Pediatr Blood Cancer. 2015:375 384. 9. Carpentieri SC, Mulhern RK, Douglas S, Hanna S, Fairclough DL. Behavioral resiliency among children surviving brain tumors: A longitudinal study. J Clin Child Psychol. 1993;22:236 246. 10. Mulhern RK, Carpentieri SC, Shema S, Stone P, Fairclough DL. Factors associated with social and behavioral problems among children recently diagnosed with brain tumor. J Pediatr Psychol. 1993;18:339 350. 11. Aarsen FK, Paquier PF, Reddingius RE, Streng IC, Arts WM, Evera-Preesman M. Functional outcome after Low- Grade Astrocytoma treatment in childhood. Cancer. 2006; 106:396 402. 12. Reddick WE, White HA, Glass JO, et al. Developmental model relating white matter volume to neurocognitive deficits in pediatric brain tumor survivors. Cancer. 2003;97:2512 2519. 13. Bonner MJ, Hardy KK, WIllard VW, Anthony KK, Hood M, Gururangan S. Social functioning and facial expression recognition in survivors of pediatric brain tumors. J Pediatr Psychol. 2008. 14. Willard VW, Hardy KK, Bonner MJ. Gender differences in facial expression recognition in survivors of pediatric brain tumors. Psychooncology. 2009;18(8):893 7. 15. McLoyd VC. Socioeconomic disadvantage and child development. Am Psychol. 1998;53:185 204. 16. Yeates KO, Swift E, Taylor HG, et al. Short- and long-term social outcomes following pediatric traumatic brain injury. J Int Neuropsychol Soc. 2004;10:412 426. 17. Anderson V, Beauchamp MH, Yeates KO, Crossley L, Hearps SJC, Catroppa C. Social competence at 6 months following childhood traumatic brain injury. J Int Neuropsychol Soc. 2013;19:539 50. 18. Moyer KH, Willard VW, Gross AM, et al. The impact of attention on social functioning in survivors of pediatric acute lymphoblastic leukemia and brain tumors. Pediatr Blood Cancer. 2012;59(7):1290 5. 19. Vannatta K, Gerhardt CA, Wells RJ, Noll RB. Intensity of CNS treatment for pediatric cancer: Prediction of social outcomes in survivors. Pediatr Blood Cancer. 2007;49:716 722. 20. Piscione PJ, Bouffet E, Mabbott DJ, Shams I, Kulkarni A V. Physical functioning in pediatric survivors of childhood posterior fossa brain tumors. Neuro Oncol. 2014;16(1):147 55. 21. Brinkman TM, Li Z, Neglia JP, et al. Restricted access to the environment and quality of life in adult survivors of childhood brain tumors. J Neurooncol. 2013;111(2):195 203. 22. Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS Med. 2010;7(7):e1000316. 23. Peterson JL, Zill N. Marital Disruption, Parent-Child Relationships, and Behavior Problems in Children. J Marriage Fam. 1986;48(2):295 307. 24. Achenbach TM. Integrative Guide to the 1991 CBCL/4-18, YSR, and TRF Profiles.. Burlington, VT: University of Vermont, Department of Psychology.; 1991. 25. Schultz KAP, Ness KK, Whitton J, et al. Behavioral and social outcomes in adolescent survivors of childhood cancer: a report from the childhood cancer survivor study. J Clin Oncol. 2007;25(24):3649 56. 26. Ellenberg L, Liu Q, Gioia G, et al. Neurocognitive status in long-term survivors of childhood CNS malignancies: a report from the Childhood Cancer Survivor Study. Neuropsychology. 2009;23(6):705 17. 27. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173 82. 28. Barrera M, Schulte F. Body Mass Index, Behavior Problems and Social Skills in Survivors of Childhood Brain Tumors. 1st Annu Conf Soc Pediatr Psychol Soc Dev Behav Pediatr. 2005.
11 29. Barrera M, Schulte F, Spiegler BJ. Factors Influencing Depressive Symptoms of Children Treated for a Brain Tumor. J Psychosoc Oncol. 2008;26(1):1 16. 30. Schulte F, Al-Khalili A, Barrera M. Impact of a pilot social skills group intervention program for survivors of childhood cancer on quality of life. Gt Lakes Reg Conf Child Heal. 2007. 31. Schulte F, Bartels U, Barrera M. A pilot study evaluating the efficacy of a group social skills program for survivors of childhood CNS tumors using a comparison group and teacher reports. Psychooncology. 2014;23(5):597 600. 32. Schulte F, Vannatta K, Barrera M. Social problem solving and social performance after a group social skills intervention for childhood brain tumor survivors. Psycho Oncology. 2014;189:183 189.